Fine Art and the Unseen Hand

Daniel Ambrosi is recognized as one of the founding creators of the emerging AI art movement and is noted for the nuanced balance he achieves in human-AI hybrid art. Ambrosi combines computational photography and artificial intelligence to create exquisitely detailed artworks that move people visually, viscerally, and cognitively. His artworks have been exhibited internationally, installed in major tech offices, featured in multiple publications, and collected by enthusiastic patrons worldwide.

Throughout the ages, a dance has taken place between art and invention; expression and technique. Traditionally, a greater appreciation and higher value has been assigned to artworks that are primarily handcrafted by their human creators, or where the hand of technology has been hidden or subdued. In those instances where the application of technology to visual expression unavoidably brought attention to itself (e.g., photography, multimedia, 3D graphics), the initial reaction of the art world has been to discredit or diminish the importance of these works to the progression of fine art. The thesis of this article is that this initial assessment contains a bias toward handmade art that is ill-founded; history has repeatedly shown that the injection of new technologies into artistic expression has never, by itself, precluded the eventual acceptance and legitimization of new art forms.

Now, advances in virtual reality and artificial intelligence are ushering in novel and powerful opportunities for visual expression where technology–more than ever–becomes an active partner in both the artistic process and the art experience. Let’s learn our lessons from the past: In this age of exponential change, it is time to recognize, accept, and even celebrate the role of technology in the evolution of fine art.

The history of art is filled with documentation and speculations regarding the tools and technologies artists used to execute their visions. Interestingly, this topic often generates heated discussions among a wide variety of scholars, not only about the plausibility of specific theories but also with respect to considerations of artistic talent and intent. Take for example the Hockney-Falco thesis advanced by artist, David Hockney and physicist, Charles M. Falco, which posited that the development of optical instruments helped advance realism in Western art since the Renaissance.

Art historians in particular seemed offended by the implication of this theory: that the Old Masters “cheated and intentionally obscured their methods.” These strong negative reactions in turn imply that these scholars believe the great works of these masters would somehow be of less value or indicative of a lesser talent if indeed these works were not produced entirely by hand and by “eyeballing it.” But why should this matter at all? Of what consequence is the artist’s manner of execution to the ultimate expression and appreciation of an art piece? If an artist was able to manifest their artistic vision fully-formed with a snap of their fingers, would that artwork have no value?

It has been said before that every great artist is a great inventor. Frustration can be a strong motivator. My own frustration with the limitations of traditional photography forced me to tinker and experiment with both old and new approaches to my art form that ultimately led to critically-acclaimed work. But that’s a story for later. First, let’s consider a more famous example where it’s highly likely (although not conclusively proved) that the interplay between art and invention led to a huge advance in artistic expression.

The 17th-century Dutch artist, Johannes Vermeer, is considered one of the greatest painters of all time; to some “an unfathomable genius.” His exquisite handling of light and near-photographic realism was many decades, if not centuries, ahead of his contemporaries. And yet very little is known of his artistic training or background. This has led to much speculation over the years among art scholars that Vermeer made use of optical aids in executing his works. This, in and of itself, isn’t particularly controversial; there are several well-established cases of artists in prior centuries using optical instruments to assist them. What is controversial–and even shocking–is the possibility raised recently by prolific inventor, Tim Jenison, that Vermeer might not have had any artistic skill or experience at all! Jenison followed this hunch to its logical conclusion by attempting to faithfully recreate a Vermeer painting from scratch despite having no artistic training whatsoever… and he succeeded:

This ambitious (and excruciatingly tedious) experiment was documented in the fascinating 2013 film, “Tim’s Vermeer” by Penn & Teller. At the conclusion of the film, Penn Jillette had this to say:

Germane to this thesis, it may be useful to consider a couple “what if” scenarios…

Giovanni Battista Piranesi was an 18th century Italian artist famous for his etchings of Rome and whose haunting depictions of imaginary prisons (“Le Carceri d’Invenzione”) continue to captivate viewers to this day:

What if, back in 1750 or so when Piranesi was working on this series, he had at his disposal a sophisticated mechanical tool that could intelligently rework his etchings line-by-line in a wide variety of ways in order to serve a specific artistic intent? For example, perhaps this device could lend an extra dreamy flow and coherence to his vision without requiring any additional effort on his part beyond configuring a few wheels and gears. Should Piranesi avoid the use of such a device for reasons of artistic purity or because the ratio of impact achieved to effort expended would be unusually high? What if an ambitious assistant demonstrated this device and showed Piranesi what it could do for him? Should that assistant be chastised and expelled from his studio for daring to suggest that the master’s work could be improved or his artistic vision more fully expressed or explored?

And what if an even more ambitious and technologically-adept assistant came up to Piranesi and said, “Hey capo!” (That’s “boss” in Italian.) “How about I take all 16 of your popular Carceri etchings and create a visual medley in three dimensions that your fans can walk through and interactively explore, perhaps while playing some mournful cello music in the background?” Would that be an extension of Piranesi’s artistic intention… or a cheap trick? You be the judge:

Let’s talk about artistic progression for a bit by focusing on a particular art form, that of representational landscapes. Currently, this art form is beginning to undergo a shift into a third paradigm. The first paradigm, landscape painting, emerged as a unique art form in 17th century Europe due in part to the popularity of evocative paintings depicting scenes in and around Rome by the French artist, Claude Lorrain. Representational landscape painting reached its pinnacle in the mid-19th century with the body of work created by the painters of the Hudson River School in the United States.

The second paradigm began with the commercial introduction of the daguerreotype camera in 1839. You can bet landscape painters of this time felt pretty threatened when–right at the top of their game–suddenly a device comes along that can capture the actual photons representing the scene. But the truth is photography was, at first, a big step down in quality and impact; the color wasn’t there, the resolution wasn’t there, you couldn’t make huge prints, etc. For these reasons, and because early photography was largely experimental, landscape photography was not initially accepted as a legitimate art form. But gradually this changed as techniques and results steadily improved and you had folks like Ansel Adams producing majestic black-and-white landscape images in the first half of the 20th century…

…and in the mid-1900’s, sensitive individuals like Eliot Porter, whose intimate landscapes did much to boost acceptance of color photography as an artistic medium…

… and in the late 20th century, folks like Peter Lik using panoramic film cameras to create large format works that, at least in some respects, surpassed those of the great representational landscape painters.

A similar progression is taking place now as landscape art enters a third paradigm with the advent of virtual reality. Again, we’re seeing a bit of a step down in quality and impact, with lots of room for improvement and plenty of problems to solve. But that will happen, and as it does, landscape VR will gradually become accepted as a legitimate art form.

So what is it exactly that drives art forward? Is artistic progression primarily driven by opportunistic and/or open-minded creators looking to capitalize on, or experiment with, new technologies and methods? Or is art mainly advanced by frustrated individuals with a burning desire to more effectively express their inner visions, thoughts, and feelings? Or is it both? And why are these artistic developments usually met with disdain or dismissal by art critics and scholars?

Consider the progression of the great J.M.W. Turner, the celebrated painter of early 19th century England. When his painting “The Dort” was first exhibited in 1818, it was hailed as “one of the most magnificent pictures ever exhibited” and Turner’s contemporary, English Romantic painter John Constable, called it the “most complete work of a genius I ever saw.”

Over time, Turner developed a looser, more evocative style that was initially met with some controversy by more conservative members of the art establishment, sometimes resulting in relegating his newer, more daring works to ancillary halls at important art exhibits, despite his fame and credibility.

When, later in life, Turner really began to push the expressiveness of his works, the push back was formidable: he was widely mocked and both his eyesight and mental health were questioned. Of course in retrospect Turner is now seen as a trailblazing genius who set the stage for the Impressionist movement that followed in the second half of the 19th century.

Turner’s progression is a vivid example of an artist pushing the limits of what his tools and technology could enable for the sake, no doubt, of more fully realizing or expressing his artistic vision. Ultimately, he was celebrated for this. Should he be any less celebrated if he had chosen to explore beyond his toolset and experimented with new technologies such as, say, photography? My point in asking that question is to illustrate that there is more than one way to skin a cat: an artist can experiment within their toolset or beyond. And both approaches should be considered equally legitimate.

I’ll now use an example of my own work where I had to do both: first pushing the limits of my existing toolset and then reaching beyond into entirely new technologies in order to achieve a specific artistic intent…

As an avid hiker, skier, and traveler, I have many times come across scenes of such breathtaking grandeur and beauty that I am struck not just visually, but also viscerally, and cognitively. When that happens, I find myself desperately wanting to bottle my experience with such fidelity that when I share it with others later, the same thing happens to them. My inability to make that happen with traditional photography was a source of endless frustration.

Ultimately, this led me to experiment with computational techniques that pushed the limits of my available toolset–within the domain of photography–to create depictions of landscapes that were more immersive, more vibrant, and significantly higher resolution than I had been able to achieve before. This was essentially accomplished by stitching and blending together dozens of individual photos in order to “force” my camera to see the world more closely to how human eyes see the world. Refinement of these methods eventually generated a body of work that took me two-thirds of the way toward achieving my goals; by all reports, I managed to impact people not just visually, but also viscerally with my new imagery.

The cognitive effect, however, remained out of my reach for the time being. Then, in July 2015, Google released a bit of open source artificial intelligence (AI) software called DeepDream that became a viral sensation. This software was initially developed as a diagnostic utility to help Google researchers understand how their own AI-based image recognition tools were working. Running DeepDream had an unexpected hallucinatory effect upon images subjected to its analysis. When the general public got their hands on this software, the Internet went crazy! People all over the world used DeepDream to turn their photos into, well… psychedelic nightmares. Like most novelties, DeepDream became an amusing fad that quickly came and went.

I, however, had other ideas in mind. It occurred to me that DeepDream might just be the tool I needed to add a level of expressiveness to my images that, like the later works of Turner and that of the Impressionists, could take my art in a new direction. Encouraged by the results and feedback on initial low-resolution experiments, I pressed on to find a way to use DeepDream on my giant images.

Unfortunately, DeepDream as released was simply not designed to operate successfully on multi-hundred megapixel images like those in my collection–it would just crash. Fortunately, I was eventually able to convince two brilliant software engineers, Joseph Smarr (Google) and Chris Lamb (NVIDIA), to modify the DeepDream source code to suit my purposes. Their turbocharged version of DeepDream made it possible to imbue my giant landscape scenes with a stunning degree of wholly unexpected form and content that is only revealed upon close-up viewing.

With the help of my ingenious engineering collaborators, I ultimately found a way to close the loop on my artistic intention and deliver a cognitive experience to my viewers. From a distance my large format artworks appear to be a photographic reality, but up close they are revealed as a digital fantasy. This forces my viewers to question the reality of what they are seeing in precisely the same way I find myself questioning the reality of what I’m seeing when I am in the presence of scenes powerful enough to affect me both visually and viscerally.

But here’s the interesting thing about this: I managed to fulfill my artistic intent with the help of a digital aid that was initially derided as a gimmick. This software has unlocked a superpower for me in the sense that I could never execute these images solely by my own hand and even trying to do so would require a prohibitive amount of time and training. And while I can set the direction of my digital aid, I’ve had to give up a degree of control in that I can’t really tell it exactly what to do and, in fact, I honestly don’t even fully understand how or why it’s doing what it’s doing. It may even be fair to say that the level of power and autonomy of my customized DeepDream software elevates this tool to that of a full-fledged artistic collaborator, albeit an artificially intelligent one. Does this diminish the value of these artworks? Does the fact that sometimes I’m sleeping while my AI collaborator tirelessly labors away, performing literally hundreds of quadrillions of math operations in the process of transmuting my images, make its contribution to my artwork worthless? I can tell you with confidence that it certainly doesn’t diminish their impact; quite the opposite is true. And besides, to paraphrase Penn Jillette’s comments above on Vermeer: It’s my composition… and (for the most part) it’s my invention.

Digital computing brings with it an ever-accelerating power to process data, reveal patterns, and otherwise extend our natural capabilities… and for this it should be celebrated. Lest you think this article is all about me, I’d like to share two more examples of other artists using computation to great effect.

Stephen Wilkes’ “Day to Night” photographic series has captured the imaginations of a wide audience and earned him a TED Talk that has been viewed over 1.5 million times as of this writing. Wilkes leverages technology to push the limits of traditional photography in order to “explore the space-time continuum” and “reveal the stories hidden in familiar locations.”

Learn more about Wilkes’ compelling work here:Capturing Both Night and Day in a Single PhotographPhotographer Stephen Wilkes has become well-known for his project titled “ Day to Night,” which features single images…petapixel.com

In the tradition of the world’s best tinkerers, Adam Magyar has engineered an ultra-high-speed photography rig that he has used to create some of the most mesmerizing video footage ever seen. His work is another vivid example of technology providing an artist a superpower with which he can express an artistic vision that would simply not be possible to execute by hand. As you’ll see in the video below, his cold unthinking camera rig ironically plumbs the depth of his unwitting subjects’ souls in a way that could only be captured surreptitiously and at lightning speed.

Learn more about Magyar’s brilliant achievements here: Einstein’s CameraHow one renegade photographer is hacking the concept of timemedium.com

Throughout history, technology–our tools and techniques–have enhanced not only our abilities and our productivity, but also our creativity. Beyond mere utility, technology can and has been leveraged toward what people want to express, not just what they want to get done. We are now reaching a point where our tools can become active partners in our creative efforts and even our collaborators, helping us to realize our artistic visions and execute our ideas in ways that we could never do on our own… but working towards goals that we set and direct. It is time to move beyond the bias toward handmade art once and for all. Doing so will not in any way diminish or threaten the value of fully handcrafted artworks moving forward, just as it should never have done to technologically-enabled art in the past.

Read more:

Daniel Ambrosi: Dreamscapes

Daniel Ambrosi: in Conversation with Serena Tabacchi